As a large part of the current data used by data scientists are created by or about humans, it is necessary and helpful to introduce to the students of data science social processes and mechanisms underlying human data. In particular, the course focuses on key issues that give rise to various challenges and/or biases in human data such as self selection, social desirability, cognitive limitations, nonindependence, contextual effects, ecological fallacy, and etc. The disposition of these issues will be inherently linked to technical issues such as sampling, measure errors, statistical control, causal inference, and etc. At the end of the course, the students are expected to be able to identify potential human errors or biases in the existing data science literature, develop systematic approaches to addressing the problems, and implement some of the suggested remedies with real world data.

Continuous Assessment: 100%